It is a common misconception — to paraphrase a commenter in the previous post on common misconceptions about publishing, that "the only two people that matter are the author and the reader (one puts creativity in, the other money: the rest add cost)".
This is a bit like saying that in commercial air travel, "the only two people that matter are the pilot and the passenger (the rest add cost)". To which I would say: what about the air traffic controllers (who stop the plane flying into other aircraft)? What about the maintenance engineers who keep it airworthy? The cabin crew, whose job is to evacuate the plane and save the passengers in event of an emergency (and keep them fed and irrigated in flight)? The airline's back-office technical support staff who're available by radio 24x7 to troubleshoot problems the pilots can't diagnose? The meteorology folks who provide weather forecasts and advise flight planners where to route their flights? The fuel tanker drivers who are responsible for making sure that the airliner has the right amount of the right type of fuel to reach its destination, and that it's clean and uncontaminated? The designers and engineers at Boeing, Airbus, Embraer, or the other manufacturers who build the bloody things in the first place ...?
I hope you can see the point I'm trying to make. To be direct: a manuscript is not a book. The author's job is to write the manuscript. The publisher's job is to turn a series of manuscripts originating from different suppliers into consistently produced books, mass-produce them, and sell them into distribution channels.
Note the phrase "series of manuscripts". Small outfits like Golden Gryphon Press (who published the first two of my Laundry novels in hardcover) work on a handful of individual titles (in GG's case, 1-3) in any given year; each book gets special attention and is handled as a separate job. Larger publishers — be they recently-graduated-from-small-press outfits like Subterranean Press or operations like Tor operate on a production line basis. Tor publish 250-300 books a year. Each incoming manuscript goes through a series of production stages, and if there's a hold-up they lose their slot in the publication queue (unless the book is very, very special, for major-bestseller values of special).
What is the production workflow for a book by a professional author working under contract with a publisher?
- The author writes a manuscript. Note: a manuscript is not a book. It's a bundle of pages of written text (or a text file on a computer: most big publishers these days insist, for no sane reason, on receiving a Microsoft Word 2003 .doc file — probably because it's the commonest format people use with support for italics, underlining, boldface, and headings). The author may suggest a title (the publisher doesn't have to accept their suggestion). The manuscript has been worked over and polished by the author to the best of their ability, but will inevitably contain typos, grammatical infelicities, continuity errors, factual errors, internal inconsistencies, and muddy bits where the author was typing on autopilot. (And these are the good, publishable manuscripts from authors with a track record. For some insight into how bad the bad stuff can get, read this and weep.)
- The author or their agent send the commissioned manuscript to their editor. (I'll write about agents and acquisitions in a future entry.) Note: outsiders often have some strange ideas about what editors do. These days, an editor is a middle manager. Their job is, in conjunction with a marketing manager, to determine what titles to acquire for their list; to negotiate deliverables and deadlines with authors: to provide high-level guidance and feedback to authors: and to ride herd on the production pipeline so that Sales and Marketing receive each quarterly or triannual batch of new titles on schedule. Editors not infrequently enjoy editing, but there's a lot more to the job than that. Here's one editor's perspective on how you wind up in the job.)
- The editor reads the MS. If it's in need of more work, they tell the author what they want doing. ("It needs a purple singing dinosaur and a surprise ending. Oh, and get rid of the DEATH guy, he's a downer.") But more often than not, they accept the manuscript. This notifies Accounting to pay the author (or their agent) the second tranche of the book advance (I'll talk about royalties, advances, and how authors make money in a future posting), the Delivery and Acquisition money. They then work out where to slot the MS into their production queue.
- Marketing. In general, the marketing spend on a book is around 5% of the dollar value of the books the publisher expects to ship. It may be more if the publisher is trying to build this particular author's sales. An editor talks about marketing, including where the money goes. Note: this activity starts when the book is delivered, before it is published — but the budget is based on expected revenue which won't be known until the orders are in. So to some extent the degree of marketing push depends on the editor's (or marketing director's) best guess as to how well the book will do.
- Scheduling. This is hammered out in the rough before the contract commissioning the book is even signed. Generally, large publishers like authors to deliver one MS per twelve months, like clockwork, because their production line is geared to turn manuscripts into published books in twelve months. However, the clock only begins ticking when the MS is accepted by the editor: so the editor looks at the calendar about 12 months ahead, works out where there's room to slot the book in — ideally in a month preceding the summer or pre-Christmas sales peak (if the book's likely to sell well) — then works backward to sort out when the various steps in the process must take place. Note that it doesn't take 12 months to produce every book — but 12 months is ample time to process the slowest book in the queue, by the recluse who doesn't use email, writes with a quill pen, and is always late reviewing the page proofs.
- Copy editing. The MS is sent to an external copy editor (usually a freelance editor specializing in the genre in question). The copy editor's job is to keep the author from embarrassing themselves in public. To this end they need to regularize spelling and typography and nomenclature, spot inconsistencies and obvious errors, fix spelling mistakes, and telepathically read the author's mind to ensure that the manuscript reflects the author's pure and original intent rather than looking as if a barrelful of monkeys had a fight for ownership of the keyboard. Traditionally, copy editors work on a paper MS using red pencils to indicate their changes; larger publishers are now switching to all-electronic workflow, hence Word change tracking and notes. The result, a Copy-Edited Manuscript or CEM, is returned to the editor within 2-4 weeks.
- The author reviews the CEM. (For some reason my publishers' editorial assistants seem to love to send me CEMs for review the week before Christmas, with a due-back deadline of January 4th. It's especially fun when two publishers pull this stunt simultaneously. Perhaps they think nobody loves me and I welcome the distraction ...?) Alas, copy editors are not, in fact, telepathic. Sometimes they get stuff wrong — rarely, they get lots of stuff wrong. So when the editor receives the CEM they bounce it to the author for review. The author can override the copy editor's changes and, if necessary, add corrections of their own. There are usually multiple changes per page — ranging from trivial stuff (serial comma policy at $PUBLISHER differs from author's usual style) to "you've referred to your hero as Joe sixteen times and Jim fourteen times in the MS — which is right?" The author typically has up to four weeks to review the CEM and return it to editorial. If it's on paper, the author must use a different colour of crayon from the copy editor, otherwise whackiness ensues. Electronic workflow involves change tracking and lots more irate marginalia, followed by hoping that Word or OpenOffice don't corrupt the file.
- Advance Reading Copies. If the book is going to be promoted heavily, the editor arranges a production run of ARCs. These are trade-paperback-sized print-outs of the manuscript, sometimes with rudimentary typesetting, and a non-glossy cardboard cover. They go to reviewers (the lead time for a book review in Publisher's Weekly, Kirkus, or another news outlet is several months), other authors (if a cover quote is being solicited), and sometimes to bookstore buyers to promote a title that's expected to sell well. Note that despite its crude appearance, an ARC costs a lot more to produce than a finished hardback — it's laser-printed and hand-bound, in small numbers, so not all books get ARCs. Someone — typically the office intern or editorial assistant — has to mail the ARCs out.
- Book design, cover design, front and back flap copy, and cover artwork. The editor pulls together a description of the book and/or the original manuscript. These are used to brief the art director, who if necessary commissions an artist to produce a painting. (Cover paintings may not be necessary if the fashion is currently for abstract/design-driven covers in marketing.) The art director also works out a suitable cover design for the book. A marketing/blurb writing person/editor also writes the flap copy (the text on the back of the book that makes you want to read it). The goal of book design is to motivate shoppers unfamiliar with the author to pick the book up. Nothing more and nothing less. (Retail psychology tells us that shoppers who handle a product are more likely to buy it. Existing loyal fans will buy it anyway. So the book design is aimed at appealing to new readers.)
- CEM delivered to Typesetting. Publishers these days own neither internal typesetting departments nor printing presses. What happens at this point is that the copy edited manuscript is sent to an external typesetting bureau, where a typesetter with a workstation running either Quark Publishing System or (increasingly, these days) Adobe InDesign sets up a Book project, imports the styles specified by the publisher's standard book design, and imports and reflows the CEM — making sure that all the changes are incorporated into the DTP file. If the book is still being produced with paper-on-pencil workflow, at this point the poor bloody typesetter has to go through it page by page entering manual corrections (and undoubtedly injecting new and creative errors of their own).
- Marketing copy. The editor develops a pitch for the book that will motivate their marketing and sales force. At a company like Tor, internal meetings are held between editorial and sales several times a year, at which the editors present their new and forthcoming titles and explain their selling points, target audience, and special angles. The sales team then go forth and push the book at store buyers (who choose which titles to stock in retail store chains) and wholesale buyers (ditto for wholesale channels). The picture at other publishers may be different; for example Ace (part of Berkley publishing group, part of Penguin, turtles, recursion, etc.) have a smaller number of higher level marketers who liaise directly with the buyers for the big chains (Barnes and Noble, Borders). This stage is critically important, because the retail and wholesale buyers place their orders before the size of the print run is finalized.
- Review page proofs. The typeset file is used to generate a PDF image of the book, as it will appear on paper. This is sent to editorial, who send a copy to the author and, hopefully, one or more proofreaders. At this point, the job of the author (and the proofreaders) is to correct errors introduced during the typesetting process, and possibly typos they didn't spot earlier — but not to make substantive changes. Again, around one month is allowed for reviewing page proofs and returning the marked-up chunk of dead trees (or an annotated PDF file, or a list of page/line number/error tuples) to editorial, who pass it back to the typesetters for fixing. (Parenthetical note: there is no such thing as a clean page proof. Authors are blind to their own persistent spelling idiosyncracies, typesetters get stuff wrong, and so on. But in general, the more eyeballs — and more proofreading passes — a set of proofs receive, the fewer errors will be left in the finished book.)
- Collate advance orders and order the print run. The sales folks have spoken to the big chain buyers and wholesalers, who guess at how many books they can sell and place orders. If these are trade books they are sold typically on 90- or 120-days net credit, sale or return. (That is: if the bookseller orders a carton of 24 books, they must either return unsold stock or pay up the agreed wholesale price when invoiced after 90 or 120 days. So they're running on credit from the publisher.) If they are mass market books they are sold on net credit, but instead of being returned for credit, the wholesaler/retailer returns the stripped covers and pulps the book block. (Which means that MMPBs unsold after 90 or 120 days are treated like spoiled grocery stock. The mass market pipeline is absolutely insane, but it mirrors how magazines are sold; it was brought in during the 1930s/1940s to sell cheap paperbacks and arguably kept the publishing industry from collapsing, but today ... let's just say I haven't met anybody who's in favour of it. Note that although mass market paperbacks are all C-format in size, not all C-format paperbacks are "mass market" books; in the UK, the sale-or-strip channel died in the late 80s/early 90s, and today — as far as I know — all British C-format paperbacks are sold on sale-or-return terms.) On the basis of these advance orders, and experience which tells them how many extra copies to print for late orders/specialist stores, the publisher orders a print run from a print shop.
- The print run: this is the moment of truth. (If the editor has coughed up a $100,000 advance against a 10% royalty on cover price, they need to ship $1M of books (cover price); if the advance orders add up to 2500 hardbacks, senior management are going to start asking difficult questions.) Some rough figures: a typical first SF/fantasy novel in hardback from a major US publisher will ship 3500-5000 copies. Anything over 10,000 is nudging towards the bestseller list; a fiction book that shifts 25,000 in its first month in hardcover is going to be in the New York Times top 20. Mass market paperback runs are much larger — 15,000-30,000 — but 25-50% will be pulped and the profit margin on them is vanishingly narrow compared to those juicy hardbacks. Oh, and the long tail applies: the top 3 on the New York Times bestseller list easily outsell the next 30 combined, and #4 to #30 combined probably outsell the next three million.
- The printing process may include multiple stages: ordering of paper (different grades of paper have different properties and costs), then printing of book blocks (containing the pages of the books), trimming, printing of colour cover flats, folding and binding, or printing of signatures, stitching and trimming, binding, printing of dust jackets, and folding (depending on whether paperback or hardback, with variations depending on whether the hardbacks are saddle-stitched — as is still common in the US market — or perfect (glue) bound — as is the case for most British hardbacks). At least the old process of cutting photographic plates is obsolete; most commercial printers these days have machines that take in a properly imposed postscript or PDF image and act (from the user's point of view) like a gigantic Postscript printer (fed by barrels of ink and giant rolls of paper). It takes 2-4 weeks from ordering the print run to the shipping pallet of finished books to appear on the publisher's warehouse loading bay, depending on how busy the print shop and how complex the manufacturing process are.
- Shipping. Someone in a warehouse has to ship out several hundred or thousand cartons of books to stores, and a smaller number of larger palletloads to wholesalers.
- Invoicing and accounting. I'm going to tip-toe past this particular pit of festering insanity. Let's just say that if you're shipping to dozens of wholesalers, a handful of large bookstore chains, and literally hundreds of small bookshops, it gets hairy fast. (Multiply by 300 titles a year with an average in-print life of 5 years if you're someone like Tor.) Note also that the contracts the publisher signs with their authors dictate that the authors will be paid a royalty (percentage share of the proceeds) based on the number of books sold and the sales channel through which they are sold. The authors almost certainly have a contractual right to order a third-party audit if they think you're screwing with their figures. And if the figures are out, you (the publisher) pay for their audit, in addition to the money you owe then.
After I hand in an MS, I expect to do another 3-6 weeks' solid work on the book before it is published — mostly in the CEM-checking and page proof-checking stages. After I hand in the MS, I expect my publisher to put in ... well, Tor produce 300 books a year with 60 staff, so it's about ten person-weeks per book in house, but this doesn't include the external copy editor, proofreader, typesetter, printer, and other outsourced tasks, which probably double it again. Overall, the process of turning a manuscript into a book is estimated to cost $7000-$20,000 — an amount comparable to the author's likely earnings from the book. In fact, the actual division of labour on a book is split roughly 50/50 between the author and the publisher.
In summary, while it's true that the author is the one with the creative input, they only do about half the work. And the other half of the job is not optional. The reason publishers exist is to provide for division of labour; if I did the other 50% to bring my rough manuscripts up to published-book-quality, I'd only be able to write half as many novels.